Abstract: Hepatocellular carcinoma（HCC） is the second deadliest cancer in China and its pathogenesis is complex. In order to seek out the core genes for poor prognosis of HCC, three gene expression profiles（GSE13471, GSE29721 and GSE6222） were found in GEO database. There were 24 HCC tissues and 17 Normal paracancer tissue in the three gene expression profiles. The differentially expressed genes（DEGs） between HCC tissues and normal tissues were identified by the GEO2 R online tool. There were total of 65 genes were consistently expressed in the three profiles, including 4 down-regulated genes and 61 up-regulated genes. The Kyoto Encyclopedia of Genes and Genomes（KEGG） pathway analysis and Gene Ontology（GO） functional annotation of DEGs was achieved by R software. Then protein-protein interaction（PPI） of these DEGs was analyzed through the Search Tool for the Retrieval of Interacting Genes（STRING） database and visualized by Cytoscape software. Then, eight hub genes had been identified on Cytohubba package in Cytoscape software and the pool prognostic value of eight DEGs expression in HCC patients were evaluated through the Kaplan-Meier plotter survival analysis tool. Furthermore, we identified three DEGs（CCNB1, CDK1 and TOP2 A） that were highly correlated with poor prognosis of HCC. For validation in Gene Expression Profiling Interactive Analysis（GEPIA）, all 3 hub genes were proved to be over-expressed in HCC samples compared to normal samples. Re-analysis of KEGG pathway revealed that two core genes（CCNB1, CDK1） were mainly enriched in p53 signaling pathway, progesterone-mediated oocyte maturation, cell cycle, oocyte meiosis, cellular senescence and human immunodeficiency virus, which were cancer-related pathways. This paper is helpful for us to have a deeper understanding of the pathophysiological mechanism of HCC and provide basis for the gene targets of subsequent drug therapy for HCC.